The recent literature, read as a case

A new paper is not a new permission. Six works that reached the desk between 2020 and 2024 are filed here at citation level. This is not a claim that each PDF was read to the last page. A number is used only as that paper stated it, under that paper's conditions.

What a review can prove

Garg, Niranjan, Prybutok, Pohlen, and Gligor, in Transportation Research Part D in 2023, reviewed 55 interdisciplinary studies of drones in the last mile. The gaps they name are these: weather, flight duration, and battery capacity are often left out; how a drone increases the receiver's access is not explained; the scene of the research leans toward the United States, and Africa and Asia are thin. A review does not erase that tilt. It puts the tilt where it can be seen.

Jazairy, Persson, Brho, von Haartman, and Hilletofth, in The International Journal of Logistics Management in 2024, read 307 articles from 2015 to 2022 from the logistics-management side. They sort twelve criteria across four parties: senders, receivers, regulators, and society. That is a map. It is not an experiment that produces the same saving in every city. Three hundred and seven is the size of the set they chose, not the number of papers in the world.

Energy refuses a single answer

Stolaroff and colleagues, in Nature Communications in 2018, argued that small delivery drones can sit below ground delivery on life-cycle energy and greenhouse gases if they are deployed with care. They also wrote that extra warehouse energy, and the longer distance flown per package, cut deeply into that advantage. Short range returns as more warehouses near the customer.

Kirschstein, in Transportation Research Part D in 2020, asked a different question. A depot-to-customer system that uses only drones consumes more energy than trucks in many settings, especially in a dense city where the truck tour is short. In a sparse rural pattern, under moderate conditions, the drone system can sit near an electric truck. The two papers do not quarrel. One says small and careful. The other says drones alone, inside urban density. The person who erases the conditions and then averages them is the person who muddles the case.

Thirty-nine customers are not a city

Roberti and Ruthmair, in Transportation Science in 2021, pushed the truck-and-drone tour to exact solutions. Their branch-and-price solves instances of up to thirty-nine customers to optimality, more than doubling the size then considered manageable. Their scenarios also say that one drone, if it is fast enough, can cut the time to finish the route by a large amount.

Thirty-nine is how far their instances reached. It is not a city's orders for a day. Wind, prohibited airspace, an aging battery, and permission are not at the center of that exact method. When a routing paper says the work gets faster, ask: faster than what, for how many customers, at what speed ratio. If the sentence has no answer, it is not yet evidence in the case.

What the Rwanda blood study measured

Nisingizwe and colleagues, in The Lancet Global Health in 2022, looked at twenty facilities in Rwanda. From 17 March 2017 to 31 December 2019, drones carried 12,733 blood-product orders. Of those, 5,517, or 43 percent, were emergencies. Mean delivery time was 49.6 minutes, with a 95 percent confidence interval from 49.1 to 50.2. They report that this was 79 minutes faster than existing road delivery on estimated driving times, and 98 minutes faster than Google Maps estimates. After drone delivery began, expirations fell by 7.1 units a month, which they translate as a 67 percent reduction at twelve months.

Observe. The comparison is not a measured road trip of the same order. It is an estimated driving time. It is not a randomized trial. Cost-effectiveness is left, by the authors, as the next question. Faster, and less waste, are records of these twenty facilities over this period. Do not carry them unchanged into another country's parcel network.